[8u/9] RFR(S): 8159244: Partially initialized string object created by C2's string concat optimization may escape

Tobias Hartmann tobias.hartmann at oracle.com
Fri Jun 17 09:52:07 UTC 2016


Hi Andrew,

sorry, I missed your message because you omitted hotspot-compiler-dev. Thanks for the clarification, I agree that it's not worth the risk.

Best regards,
Tobias

On 14.06.2016 12:41, Andrew Haley wrote:
> On 14/06/16 11:08, Tobias Hartmann wrote:
>> So I wondered why a StoreStoreBarrier is not sufficient in this case.
> 
> It probably is, but I'm not certain that we can rely on it, and
> it's not much of an efficiency gain for the risk.  We already know
> that there are no write-only objects in HotSpot because of locking.
> 
> Andrew.
> 


More information about the jdk8u-dev mailing list